Extend analyze.py for ISMCTS metrics

- Loss section: add ISMCTS policy/value twin-axis panel alongside the
  Deep CFR advantage/strategy panel. Both render only when their keys
  exist; the unused side shows "No data".
- Memory Size: add memory/replay alongside memory/advantage and
  memory/strategy.
- New MCTS section (analysis_09_mcts.png): visit-count entropy, value
  prediction error, policy/MCTS KL — three iter-time scalars emitted by
  IsMctsTrainer. Auto-skips on Deep CFR runs (no data).

src/coolrl_lost_cities/games/classic/deep_cfr/analyze.py

@@ -90,6 +90,14 @@ SECTIONS: tuple[SectionSpec, ...] = (
                 secondary_metrics=("loss/strategy",),
                 secondary_ylabel="strategy CE",
             ),
+            PlotSpec(
+                "Losses (ISMCTS policy / value)",
+                ("loss/policy",),
+                "policy CE",
+                kind="train",
+                secondary_metrics=("loss/value",),
+                secondary_ylabel="value MSE",
+            ),
             PlotSpec(
                 "Samples",
                 ("samples/advantage", "samples/strategy"),
@@ -98,7 +106,7 @@ SECTIONS: tuple[SectionSpec, ...] = (
             ),
             PlotSpec(
                 "Memory Size",
-                ("memory/advantage", "memory/strategy"),
+                ("memory/advantage", "memory/strategy", "memory/replay"),
                 "samples",
                 kind="train",
             ),
@@ -287,6 +295,30 @@ SECTIONS: tuple[SectionSpec, ...] = (
             ),
         ),
     ),
+    SectionSpec(
+        "MCTS",
+        "analysis_09_mcts.png",
+        (
+            PlotSpec(
+                "Visit-count entropy at root",
+                ("mcts/avg_visit_entropy",),
+                "nats",
+                kind="train",
+            ),
+            PlotSpec(
+                "Value prediction error",
+                ("mcts/value_prediction_error",),
+                "MSE (score units)",
+                kind="train",
+            ),
+            PlotSpec(
+                "Policy / MCTS KL",
+                ("mcts/policy_mcts_kl",),
+                "KL (nats)",
+                kind="train",
+            ),
+        ),
+    ),
 )
 
 SUMMARY_EVAL_METRICS: tuple[tuple[str, str, float], ...] = (